444 Mr, O. Thomas on 



America. But tliree of them — suhUneatu.i, ferru^irieufi, nnd 

 dornalis — extend into E;isteru Brazil, the fir.st-named in tlie 

 nortii, the other two in the south. T./errufjineus and dorsalis 

 are also exceptional in j)ossessing 2 — 2 = 8 mamrage. 



No doubt other described species will be found to be refer- 

 able to T//omasomi/s, but the above are all that I have as yet 

 been able to identify. 



But, further, a study of the same mammary and palatal 

 characters in the series of forms that have been allocated to 

 llldpidomys brings out the fact that among these there are 

 two distinct groups differing from eacli other exactly as do 

 Oryzomys and Thomasomys ; for up to the present no accurate 

 definition of " Rhijjidomys " has been given, and the fact that 

 the tail of any species is more or less tufted and that the 

 animal had certain other external peculiarities have been 

 taken as sufficient reason for its reference to what I now 

 find to be the composite genus Rhipidomys, 



The true Rhijjidomya has 1 — 2 = 6 mammte, a long heavily 

 tufted tail, broad climbing feet, and the })alatal characters of 

 Thomasomys, from which it is to be distinguished by its 

 external peculiarities and by the presence of well-marked 

 divergent supraorbital ridges^ these being practically absent 

 in Thomasomys. 



The species to which the following specific names have 

 been given appear to belong to this genus : — leucodactylus, 

 Tschudi (type of genus) ; mastucalisj Lund ; macrurus, 

 Gerv. ; latt'nianus, Tomes; ochrogaster and Couesi, Allen; 

 tScIateri, Goodfelloici, venezuehe, venustus, microtis, pictor, 

 nitela, Sindfulviventer, Thos.; and MacconneUi, de Wintou ; 

 but ihe last-named is somewhat aberrant in other characters 

 than those above mentioned, and may hereafter prove to be 

 sc])arable from the rest. 



The species belonging to the second group have absolutely 

 the skull of Oryzomys, and they also, so far as is known, 

 have 2 — 2 = 8 niamma3. It is clear, therefore, that they 

 should be altogether removed from Rhipidomys, and either 

 assigned to Oryzomys or form a special group of their own. 



On the whole it appears to me they may best be regarded 

 as a subgenus of Oryzomys^ as follows : — 



CECOMYS*, subgen. nov. 



Number of niammK (2 — 2 = 8) and essential skull-characters 

 as in Oryzomys, though there is a tendency for the brain-case 



* oiKoi, a house. Quite a number of spocimeus, of different species, 

 are noted as haviiijr been canjrht in native houses. 



